RedactionKeyword
RedactionKeyword
enumerates sensitive words or phrases detected on the drawing that may need masking before documents are shared. Each instance records the matched term, its location, and confidence so downstream workflows can decide whether to automatically redact, replace, or flag the text for review. The model also stores surrounding context and document references, helping reviewers judge whether a match is business-critical or incidental. Use the object to enforce confidentiality policies, comply with ITAR or NDA requirements, and protect customer names or proprietary part numbers while still leveraging the rest of the extracted data. Because the keyword list is structured, you can integrate it with ticketing systems, drive audit logs, or build dashboards highlighting recurring sensitive terminology.